WIXLIB

Shi et al. J Exp Clin Cancer Res(2022) 41:234Page 3 of 17The AASEs in OABLs were identified using rMATS[15]. All the sequences and annotations used in thisanalysis were based on GRCh38 genome assembly. AnASE with a ΔInclevel value between the OABLs andcontrols of more than 5% ( ΔInclevel 0.05) and adjp-value of 0.01 was identified as an AASE. The list ofannotated splicing factors and regulators was downloaded from the SpliceAid-F database and the study byNostrand et al. [16, 17] (Supplementary Table S9).pLKO.1-shADAR#2:CCG G GC CCA C TG T TA TCT TCA C TT TCT CGA GAA AGT GAA GAT A AC AGT G GG CTT T TTG) were purchased from Shanghai Genomditech (Shanghai, China). Non-targeting shRNA wasused as the control. The lentiviral vectors and packaging vectors were transfected into the 293 T packagingcell line using the PolyJet In Vitro DNA TransfectionReagent (SignaGen). Retroviral vectors were transfected. Targeted cells were infected with lentivirus inthe presence of 8 μg/ml polybrene (Sigma).Differential expression analysisWestern blotFor transcriptomic data, low-abundance transcriptswere removed. For proteomic data, low-abundanceproteins and proteins missing in 20% patientswere removed, and K-nearest neighbor imputationmethod was used to complete values in proteomicdata for proteins missing in 20% patients. Differential expression analysis was performed using limmaR package (version: 3.46.0) [18] after normalization.We set log2(foldchange) log2(1.5) and p-value 0.05 as the threshold for transcriptome data, and log2(foldchange) log2(1.2) and p-value 0.05 as thethreshold for proteome data.Cells were lysed in mammalian protein extraction reagent (Pierce). After protein quantification using a bicinchoninic acid protein assay kit (Pierce), 60 μg of totalprotein was separated by 10% SDS-PAGE under denaturing conditions and transferred to PVDF membranes(Millipore). Membranes were blocked in 5% non-fat milk(Bio-Rad) and then incubated with primary antibodies,followed by incubation with a secondary antibody conjugated with horseradish peroxidase (1:10,000; AmershamBiosciences). Immunoreactive proteins were visualizedusing the LumiGLO chemiluminescent substrate (CellSignaling Technology). The primary antibodies used inthis study are as follows: b-Actin (1:10,000; Sigma); andADAR1 (1:1000; Cell Signaling Technology).AASE identification and analysisFunctional enrichment analysisThe Metascape tool (http:// metas cape. org) [19] wasused to identify biologic genesets of the selected genes.Gene Set Enrichment Analysis (GSEA, https:// w ww. gsea- msigdb. org) [20, 21] and Gene Set Variation Analysis (GSVA) [22] were performed to investigate deregulated genesets between groups. Particular MolecularSignatures Database collections were included in theGSEA and GSVA analyses. Hallmark (H), curatedcanonical pathways (C2:CP), and gene ontology BP (C5:GO BP) gene set collections were included to identifygenesets.Cells and cell cultureNHL cell lines were provided by Cell Bank and StemCell Bank, Chinese Academy of Sciences. The Rajiand SU-DHL-4 lymphoma cell lines were cultured inRPMI1640 (Invitrogen, CA) supplemented with 10%fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco) and maintained at 37 C in a 5% CO2humidified atmosphere.Virus transduction and generation of stable cell linesTwo individual lentiviral vectors containing shRNAs targeting human ADAR (pLKO.1-shADAR#1:CCG G CG G AT A C T A C A C CC ATC C AT T C T C GA G A A TG G ATG G G T G TA G TA TC C G T T T T TG ;Cell proliferation assayTo assess cell proliferation, a Cell Counting Kit-8 (CCK-8,New Cell & Molecular Biotech, China) was used following the manufacturer’s instructions. Cells were seededinto a 96-well plate at 1   103 cells per well with 100 μlmedium and cultured at 37 C with 5% CO2. CCK8 solution was added (10 μl per well) and cells were incubatedfor 3 h before measuring the absorbance at 450 nm.Immunohistochemistry (IHC) and immunofluorescent (IF)stainingIHC and IF were performed following standard procedures. Five-micron thick formalin-fixed paraffin embedded (FFPE) human tissue sections were used for theexperiments. Stained slides were digitized using the Pannoramic DESK (3D HISTECH) with a 40 objective lens.For IF, sections were stained with the Djc9 antibody(#NBP1–87903, Novusbio, 1:50) and CD20 antibody(#GB14030, Servicebio, 1:200) following the manufacturer’s instructions. Mean fluorescent intensity (MFI)was quantified using ImageJ (https:// imagej. nih. gov/ ij/). Colocalization analysis was performed using ImageJplugin JACoP.For IHC, sections were stained with the Djc9 antibody (#NBP1–87903, Novusbio, 1:500). Staining intensity score, nuclear staining intensity score, and staining

Shi et al. J Exp Clin Cancer Res(2022) 41:234percentage were quantified using ImageJ IHC Profilerplugin. Staining index was calculated as follows: staining intensity score (0–3) staining percentage (0–100%).Staining intensity score was graded as follows: score 0:negative staining, score 1: weak staining, score 2: moderate staining, score 3: strong staining.DefinitionsLocal recurrence was defined as lymphoma recurrenceat the orbital region. Distant recurrence was defined aslymphoma recurrence at an extra orbital site that wasnot initially involved. Progression-free survival (PFS)was calculated from the date of diagnosis to recurrence,progression, death, or the most recent follow-up. Overallsurvival (OS) was defined as the time from initial diseasediagnosis to death by any cause or until the most recentfollow-up. Recurrence-free survival (RFS), local recurrence-free survival (LRFS), and distant recurrence-freesurvival (DRFS) were calculated from the date of initialtreatment to the corresponding recurrence, or until themost recent follow-up.Statistical analysisAll statistical tests were two-sided and P-values 0.05were considered statistically significant. Statistical analyses were performed using the R software developed bythe R Development Core Team at R Bioconductor [23]and GraphPad Prism /).For comparison of continuous variables, we used theMann–Whitney U test for correlation results as previously mentioned [11], and the Student’s t-test for others. For correlation analysis, we applied the Spearmanrank correlation test for global protein-mRNA correlation following previous observations [24], and Pearson’scorrelation test for other continuous variables. Survivalcurves were generated using the Kaplan–Meier methodand compared by log-rank test. Cox proportional hazardsmodels were constructed to identify prognostic factorsfor OABL. Adjusted hazard ratios with 95% confidenceintervals were calculated. Logistic regression modelsand lasso regression models were constructed to identifydiagnostic markers for OABL.ResultsIntegrated proteotranscriptomic landscape of OABLTo obtain a comprehensive view of the molecular characteristics of OABLs, we performed an integratedproteomic and transcriptomic analysis of 40 OABLs,including 28 EMZLs, 8 DLBCLs, 2 MCLs, and 2 smalllymphocytic lymphomas (SLLs), and 31 control specimens including 12 IOIs, 6 RLHs, and 13 normal orbitaltissues (Supplementary Table S1, Fig. 1A). Because of thePage 4 of 17contamination of plasma proteins, TMT labeled liquidchromatography-mass spectrometry was not performedin one EMZL, three IOI, and two normal samples. Therefore, after a quality control process, we obtained proteomic data of 39 OABLs and 26 controls (SupplementaryTable S2). Because of RNA degradation, RNA-seq wasnot performed in 7 EMZLs, 1 DLBCL, 1 SLL, 4 IOIs, 3RLHs, and 7 normal tissues. After quality control, weobtained transcriptomic data of 31 OABLs and 17 controls (Supplementary Table S3).By performing hierarchical clustering of highly variable genes (HVGs) and samples in transcriptomic andproteomic data, we found that most of the OABLs andcontrols were divided into two groups. While OABLand normal specimens were clearly divided, some of theinflammation samples, especially RLH samples, weremisclassified into the OABL group (Supplementary Fig.S1). These results validated our sampling and analysisprocesses and echoed the observed association betweeninflammation and lymphoma [25–28].After removal of low-abundance transcripts/proteinsand completion of missing values, we mapped transcriptsand proteins to Ensembl IDs and detected 3639 proteinmRNA pairs in these samples. For these pairs, we calculated differentially expressed genes (DEGs) betweenOABLs and controls (Fig. 1B, Supplementary Table S4).The results showed that 787 genes were upregulatedand 542 genes were downregulated in the proteome ofOABLs compared with controls. Eight hundred six geneswere upregulated and 445 genes were downregulated inthe transcriptome of OABLs compared with controls.After matching results of omics data, we found that 451genes were concordantly upregulated (CO-UP), 386genes were concordantly downregulated (CO-DOWN),and 19 genes were discordantly dysregulated. The CO-UPDEGs were mainly enriched in immune-related, GTPasesignaling, negative regulation of phosphate metabolicprocess and DNA recombination terms. The CO-DOWNDEGs were mainly enriched in normal tissue development and organization, monocarboxylic acid, and arginine and proline metabolism terms (Fig. 1C).To avoid reported potential systematic biases [10], wefurther performed GSEA to identify overlapping dysregulated gene sets in OABLs. A total of 1023 gene sets werecommonly identified by proteomic and transcriptomicdata (Supplementary Table S5). Among these, 763 genesets were significantly dysregulated in at least one typeof omics (FDR 0.2), and 725 were concordantly dysregulated and 38 were discordantly regulated (Fig. 1D).Arranged by the sum of NES rank in each omic, CO-UPgene sets were mainly enriched in mRNA processing andsplicing, DNA damage and repair, and protein sumoylation. CO-DOWN gene sets were mainly enriched in

Shi et al. J Exp Clin Cancer Res(2022) 41:234Page 5 of 17Fig. 1 Proteotranscriptomic characteristics of OABL. A Sampling workflow of transcriptomic and proteomic cohort. B Scatterplot of differentiallyexpressed genes (DEGs) of 3639 protein-mRNA pairs identified in the transcriptomic cohort (31 OABLs and 17 controls) and the proteomic cohort(39 OABLs and 26 controls) between OABLs and controls. The horizontal line is at proteomic log2(FC) log2(1.2); vertical line is at transcriptomic log2(FC) log2(1.5). Genes with p 0.05 are labeled as black. C Bar plot of top 20 enrichment terms identified in concordant upregulated anddownregulated genes. D Bubble plot of differentially enriched gene sets identified by GSEA in the transcriptomic cohort (31 OABLs and 17 controls)and the proteomic cohort (39 OABLs and 26 controls) between OABLs and controls. Gene sets with FDR 0.2 in at least one of transcriptomic andproteomic analysis are displayed. E Bubble plot of top 15 up/down regulated gene sets identified by GSEA

Shi et al. J Exp Clin Cancer Res(2022) 41:234normal tissue development and organization, and aerobic glucose metabolism (Fig. 1E). As our study containedmultiple subtypes of OABL, we performed GSVA inthe proteomic data and examined the robustness of thedysregulations (Supplementary Fig. S2A, Supplementary Table S6, Supplementary Methods). The variationsof these top-ranked genesets were consistent among allfour subtypes. For discordantly dysregulated genes andpathways, immune, Golgi traffic and amide metabolismrelated gene sets were identified by DEG enrichment andGSEA analyses (Supplementary Fig. S2B, C).Global protein‑mRNA concordance is a distantrecurrence‑related characteristic of OABLNext, we examined the relationship between proteinand mRNA abundance and its association with diseasecharacteristics. Global protein-mRNA concordance wascomputed as the Spearman correlation result of all pairedprotein and mRNA abundance in each patient (Fig. 2A).We analyzed this concordance in patients with both proteomic and transcriptomic data (Fig. 1A). Consideringthe different transcript/protein abundance distributionbetween OABLs and controls, we analyzed proteinmRNA pairs separately in these two groups. We identified 3818 protein-mRNA pairs in OABLs and 3728 pairsin controls. The concordance was significantly higher inOABLs (median rho 0.364) than in controls (medianrho 0.208, p 0.01, Fig. 2B-C). Considering the association between inflammation and lymphoma [25–28], wecompared the global concordance across subgroup specimens. The results showed that OABLs exhibited a relatively higher concordance than other groups (median rhoof RLH 0.254, IOI 0.23, normal 0.16). These dataindicated that the increased correlation between proteinand mRNA abundance is a disease-associated characteristic of OABL.We next analyzed if the concordance is associatedwith disease aggressiveness. First, we compared theconcordance across subtypes, Ann Arbor stage, andprognostic risk factors (Fig. 2D). The concordancewithin no-EMZL subtypes was significantly higherthan that of EMZL (p 0.039), an indolent lymphomasubtype. High concordance was significantly associated with higher LDH (p 0.014) and IPI (p 0.046),Page 6 of 17two prognostic risk factors of NHL. High concordance was relatively associated with a higher Ann Arborstage (p 0.054) (Supplementary Fig. S3A). Next, weevaluated the correlation between proliferation abilityand global concordance (Fig. 2E). The result showed astrong positive association between Ki67 protein abundance and global concordance in OABLs (r 0.495,p 0.005), but this association was not present in thecontrols (r 0.203, p 0.527). Hence, higher globalconcordance was associated with disease aggressiveness solely in OABL.We then tested the correlation between the proteinmRNA concordance and OABL prognosis. We dividedOABL patients into two groups using the median valueof concordance and compared PFS, OS, RFS, LRFS, andDRFS between the two groups (Fig. 2G, SupplementaryFig. S3B). Among the 15 patients in the high rho group,6 showed recurrence, 1 showed local recurrence, and 6showed distant recurrence. For the 15 patients in the lowrho group, 2 showed recurrence, 1 showed local recurrence, and 1 showed distant recurrence (Supplementary Table S1). Survival analysis revealed that a globallyincreased concordance in OABL was significantly associated with reduced DRFS (p 0.037) and relatively associated with reduced RFS (p 0.083), but not with PFS(p 0.19) or OS (p 0.26).We then compared the global concordance betweenpatients with and without the recurrence events (Fig. 2F,Supplementary Fig. S3C). High concordance was significantly associated with distant recurrence events(p 0.0034) and recurrence events (p 0.0072), but notwith local recurrence events (p 0.41). We analyzed therelationship between the global concordance and recurrence in small B-cell lymphoma (SBL), EMZL, DLBCL,and other subtypes to ensure the robustness of the finding (Supplementary Fig. S3C). High concordance wassignificantly associated with distant recurrence events inSBL (p 0.0059) and EMZL (p 0.039). Despite the lowincidence of recurrence and limited number of patients,the median value of patients with distant recurrencewas still higher than that of patients without the eventsin DLBCL and other subtypes. These data demonstratedthat the high global protein-mRNA concordance was apredictive factor for distant recurrence in OABL.(See figure on next page.)Fig. 2 The match-subject analysis identifies global protein-mRNA concordance as an OABL-associated characteristic. A Schematic diagram ofglobal protein-mRNA concordance calculation. B Density plot showing the global Spearman correlation for protein-mRNA pairs within OABLs(n 3818 protein-mRNA pairs) and controls (n 3728 pairs). C Concordance of protein-mRNA pairs is significantly higher in OABLs compared withthe control or normal group and relatively higher compared with the RLP or IOI group. D Global protein-mRNA concordance is associated withprognostic risk factors. No-EMZL subtype, high LDH, and high IPI score have an increased concordance. E Global protein-mRNA concordance ispositively correlated with the MKI67 proteomic abundance in the OABLs (r 0.495, p 0.005) but not in the controls (r 0.203, p 0.527). Blueline shows liner regression. F High global protein-mRNA concordance is associated with distant recurrence in OABL. G Kaplan-Meier plot showshigh global concordance in OABLs is associated with decreased distant recurrence-free survival. H Bar plot of top 20 gene sets identified by GSVAcorrelated with global protein-mRNA concordance

Shi et al. J Exp Clin Cancer Res(2022) 41:234Page 7 of 17Fig. 2 (See legend on previous page.)Next, we investigated the potential regulators andbiological implications of the abnormally upregulatedglobal protein-mRNA concordance. Because the globalconcordance is an intrinsic continuous variable, weexamined the correlation between GSVA results andthe concordance (Fig. 2H). In the top 20 positively correlated genesets, 8 gene sets were TP53-related gene sets,and the others were mostly immune-related gene sets.

Shi et al. J Exp Clin Cancer Res(2022) 41:234ECM-associated gene sets accounted for the majority ofnegatively correlated gene sets.These findings indicate that increased global proteinmRNA concordance is a novel molecular characteristicof OABL that is associated with disease aggressivenessand higher risk of recurrence. This abnormally upregulated concordance in OABL is positively related to theTP53 pathway.Trend analyses identify alternative splicingas an inflammation‑independent signature of OABLIn the proteotranscriptomic data, we observed a similarity of molecular characteristics between inflammation and OABL samples through hierarchical clustering,principle component analysis, and global protein-mRNAconcordance (Fig. 2C, Fig. 3A-B, Supplementary Fig. 1).As previous studies demonstrated the activation of NFκBsignaling pathway in both inflammation and NHL [25,26], we performed hierarchical clustering in the NFκBsignaling pathway across subgroups (Fig. 3C). The abundance of NFκB-related genes progressively increasedfrom normals to inflammations, and to OABLs, whichwas consistent with the previous reports [25, 26]. However, issues remained as: what extent the similarity is;which pathways discriminate OABL from inflammation;and whether these pathways are driver events of OABL.To address these questions, we constructed a robustinflammation-OABL signature in proteomic data bysupervised and unsupervised clustering genes across thenormal, inflammation, and OABL groups. First, we performed the t-test of protein abundance between eachtwo groups and hypothesized nine dysregulated patterns of genes (Fig. 3F, Supplementary Methods). Mostwere constituted by the upregulated patterns (cluster3u24.6%; cluster2u, 21.94%; cluster4u, 11.68%). Interestingly, in these upregulated patterns, inflammations couldnot be discriminated from OABLs in a majority of genes(906/2144, 42.3%). We additionally performed an unsupervised k-means clustering for normalized HVGs (geneswith top 50% MAD, k 4) (Fig. 3D-E). Cluster numberswere determined by the elbow plot (Supplementary Fig.S4A).Page 8 of 17Combining the results of k-means clustering andt-test gene patterns, we identified five clusters of genes:inflammation mimic upregulated genes (MIMIC-UP),vaguely upregulated genes (VAGUE-UP), OABL specificupregulated genes (SPECIFIC-UP), inflammation mimicsdownregulated genes (MIMIC-DOWN), and OABL specific downregulated genes (SPECIFIC-DOWN) (Fig. 3G,Supplementary Table S7). Because upregulated genesconstituted most of the clustered proteins, we focusedon MIMIC-UP and VAGUE-UP, which representedextremely different patterns of dysregulation. MIMICUPs were mostly enriched in immune-related gene sets,while SPECIFIC-UPs were mostly enriched in gene setsthat related to mRNA metabolism and splicing, DNAdamage and metabolism, and chromatin remodeling(Fig. 3H).These results clearly demonstrated that the similarity between inflammation and OABL is not only in theNFκB pathway but also in a larger immune landscape.More importantly, we identified gene sets specificallydysregulated in OABL, including mRNA splicing andwell-known pathways associated with malignancy development (DNA damage, chromatin remodeling).Alternative splicing and its regulators potentially influenceOABL development and progressionAltern

To assess cell proliferation, a Cell Counting Kit-8 (CCK-8, New Cell & Molecular Biotech, China) was used follow-ing the manufacturer's instructions. Cells were seeded into a 96-well plate at 1 10 3 cells per well with 100 μl medium and cultured at 37 C with 5% CO 2. CCK8 solu-tion was added (10μl per well) and cells were incubated